在 tf.keras 中正确设置 GAN 实现中的 .trainable 变量

Question

我对GAN 实施中.trainable的声明感到困惑。tf.keras.model

给出以下代码片段（取自此 repo）：

class GAN():

    def __init__(self):

        ...

        # Build and compile the discriminator
        self.discriminator = self.build_discriminator()
        self.discriminator.compile(loss='binary_crossentropy',
            optimizer=optimizer,
            metrics=['accuracy'])

        # Build the generator
        self.generator = self.build_generator()

        # The generator takes noise as input and generates imgs
        z = Input(shape=(self.latent_dim,))
        img = self.generator(z)

        # For the combined model we will only train the generator
        self.discriminator.trainable = False

        # The discriminator takes generated images as input and determines validity
        validity = self.discriminator(img)

        # The combined model  (stacked generator and discriminator)
        # Trains the generator to fool the discriminator
        self.combined = Model(z, validity)
        self.combined.compile(loss='binary_crossentropy', optimizer=optimizer)

    def build_generator(self):

        ...

        return Model(noise, img)

    def build_discriminator(self):

        ...

        return Model(img, validity)

    def train(self, epochs, batch_size=128, sample_interval=50):

        # Load the dataset
        (X_train, _), (_, _) = mnist.load_data()

        # Adversarial ground truths
        valid = np.ones((batch_size, 1))
        fake = np.zeros((batch_size, 1))

        for epoch in range(epochs):

            # ---------------------
            #  Train Discriminator
            # ---------------------

            # Select a random batch of images
            idx = np.random.randint(0, X_train.shape[0], batch_size)
            imgs = X_train[idx]

            noise = np.random.normal(0, 1, (batch_size, self.latent_dim))

            # Generate a batch of new images
            gen_imgs = self.generator.predict(noise)

            # Train the discriminator
            d_loss_real = self.discriminator.train_on_batch(imgs, valid)
            d_loss_fake = self.discriminator.train_on_batch(gen_imgs, fake)
            d_loss = 0.5 * np.add(d_loss_real, d_loss_fake)

            # ---------------------
            #  Train Generator
            # ---------------------

            noise = np.random.normal(0, 1, (batch_size, self.latent_dim))

            # Train the generator (to have the discriminator label samples as valid)
            g_loss = self.combined.train_on_batch(noise, valid)

在模型定义期间，self.combined判别器的权重被设置为self.discriminator.trainable = False但从未重新打开。

尽管如此，在训练循环期间，判别器的权重将会改变：

# Train the discriminator
d_loss_real = self.discriminator.train_on_batch(imgs, valid)
d_loss_fake = self.discriminator.train_on_batch(gen_imgs, fake)
d_loss = 0.5 * np.add(d_loss_real, d_loss_fake)

并将在以下期间保持不变：

# Train the generator (to have the discriminator label samples as valid)
g_loss = self.combined.train_on_batch(noise, valid)

这是我没想到的。

当然，这是训练 GAN 的正确（迭代）方法，但我不明白为什么我们在对self.discriminator.trainable = True判别器进行一些训练之前不必通过。

如果有人对此有解释，那就太好了，我想这是理解的关键点。

Answer 1

当您对 github 存储库中的代码有疑问时，检查问题（开放的和已关闭的）通常是个好主意。 此问题解释了为什么该标志设置为False。它说，

由于self.discriminator.trainable = False是在判别器编译后设置的，因此不会影响判别器的训练。然而，由于它是在编译组合模型之前设置的，因此在训练组合模型时鉴别器层将被冻结。

还讨论了冻结 keras 层。